Automatic config generation for automatic delay cal selection

[tikk3r]

No description provided.

[tikk3r]

Core functionality all seems to work now. Just stuck on why the generate_skymodels step is still executed even though the array length checked in the when condition should be >0.

[tikk3r]

why the generate_skymodels step is still executed even though the array length checked in the when condition should be >0.

This looks like it's maybe just a Toil oddity.

Functionality all works, but the vlass starting models are not used it seems. A manual facetselfcal run against the same VLASS image yields wildly different astrometry even if I completely remove the generat_skymodels step from the CWL code as a test, so I suspect the pipeline is still generating a point source and passing that.

Left to right: VLASS image, pipeline image 9, manual image 3

[tikk3r]

It looks like the difference may be in the PyBDSF commands for source finding. The facetselfcal-generated model has more components than what skynet outputs:

Skynet

# (Name, Type, Patch, Ra, Dec, I, Q, U, V, MajorAxis, MinorAxis, Orientation, ReferenceFrequency='144e+06', SpectralIndex='[]', LogarithmicSI) = format
, , P0, 00:00:00, +00.00.00
ME0,GAUSSIAN,P0,10:42:07.07427757,+56.20.35.11342256,0.7501937174536204,0.0,0.0,0.0,1.6648132122494166,1.2344896254335085,109.43104961290301,144e+06,[-0.7],true
ME1,GAUSSIAN,P0,10:42:06.08426152,+56.20.42.68432263,0.0933097437405896,0.0,0.0,0.0,1.6331785957898932,0.34046779089727963,136.72680994037083,144e+06,[-0.7],true

Facetselfcal

format = Name, Type, Patch, Ra, Dec, I, Q, U, V, MajorAxis, MinorAxis, Orientation, ReferenceFrequency='3.00000e+09', SpectralIndex='[]'

, , ILTJ104206.59+562039.5_vlass_patch_s0, 00:00:00, +00.00.00
ILTJ104206.59+562039.5_vlass_w0_i0_s0_g0, POINT, ILTJ104206.59+562039.5_vlass_patch_s0, 10:42:7.119336, +56.20.35.033496, 2.555e-02, 0.0, 0.0, 0.0, 0.00000e+00, 0.00000e+00, 0.00000e+00, 3.00000e+09, [-0.8]
ILTJ104206.59+562039.5_vlass_w0_i0_s0_g2, POINT, ILTJ104206.59+562039.5_vlass_patch_s0, 10:42:6.991234, +56.20.35.438892, 1.120e-02, 0.0, 0.0, 0.0, 0.00000e+00, 0.00000e+00, 0.00000e+00, 3.00000e+09, [-0.8]
ILTJ104206.59+562039.5_vlass_w0_i0_s0_g3, GAUSSIAN, ILTJ104206.59+562039.5_vlass_patch_s0, 10:42:6.768397, +56.20.35.979193, 6.761e-03, 0.0, 0.0, 0.0, 4.41003e+00, 0.00000e+00, 1.68543e+02, 3.00000e+09, [-0.8]
, , ILTJ104206.59+562039.5_vlass_patch_s1, 00:00:00, +00.00.00
ILTJ104206.59+562039.5_vlass_w0_i0_s1_g1, POINT, ILTJ104206.59+562039.5_vlass_patch_s1, 10:42:6.086021, +56.20.42.687386, 3.819e-03, 0.0, 0.0, 0.0, 0.00000e+00, 0.00000e+00, 0.00000e+00, 3.00000e+09, [-0.8]
, , ILTJ104206.59+562039.5_vlass_patch_s2, 00:00:00, +00.00.00
ILTJ104206.59+562039.5_vlass_w0_i1_s2_g4, GAUSSIAN, ILTJ104206.59+562039.5_vlass_patch_s2, 10:42:4.669273, +56.21.1.664814, 2.484e-03, 0.0, 0.0, 0.0, 2.34113e+00, 9.84720e-01, 6.75894e+01, 3.00000e+09, [-0.8]

[tikk3r]

After some digging around in skynet.py it is indeed related to how it manipulates the sky model found by PyBDSF. Writing out the PyBDSF model directly results in the same model as what facetselfcal generates.

This means as far as I can tell the goal of this branch is there, so I'll mark it ready for review and we can fix skynet in another one.